Tall, slender and with unkempt hair, Arthur Mensch arrived in Paris' vast tech hub for a speech last month wearing jeans and a bicycle helmet. He had a reserved demeanor befitting someone European authorities are counting on to help take the region into a high-stakes contest with the United States and China over his artificial intelligence.
Mensch, 31, is the CEO and founder of Mistral and is considered by many to be one of the most likely challengers to OpenAI and Google. “You have become a symbol of AI in France,” British investor Matt Clifford said on stage.
Mensch's company shot into the spotlight just a year after it was founded in Paris with two college friends. As Europe races to grab a foothold in the AI revolution, the French government has named Mistral as its best hope for creating a standards company and is lobbying EU policymakers to ensure the company's success. Ta.
Artificial intelligence will be rapidly integrated into the global economy over the next decade, but European policymakers and business leaders worry that growth and competitiveness will suffer if the region does not catch up. are doing. Their concerns stem from a belief that AI should not be dominated by big tech companies like Microsoft and Google, which could create global standards that conflict with other countries' cultures and politics. At issue is a larger question: which artificial intelligence models will end up impacting the world, and how should they be regulated?
“The problem with not having a European champion is that the roadmap is set by the United States,” Mensch said. As recently as 18 months ago, he was working as an engineer at Google's DeepMind lab in Paris, building AI models. His co-founders, Timothy Lacroix and Guillaume Lampre, are also in their 30s and held similar positions at Meta.
In an interview in Mistral's modest, whitewashed offices on Canal Saint-Martin in Paris, Mensch said the U.S. tech giant set the ground rules for powerful new technologies that will affect millions of lives. “It's not safe to trust,” he said.
“You can't have strategic dependencies,” he said. “That's why we want to create European champions.”
Europe has struggled to produce meaningful technology companies since the dot-com boom. According to a report by France's Artificial Intelligence Commission, the United States has produced Google, Meta, and Amazon, and China has produced ByteDance, which owns Alibaba, Huawei, and TikTok, but Europe's digital economy is underperforming. I couldn't raise it. The 15-member commission, including Mensch, warned that Europe was lagging behind on AI, but said Europe could take the lead.
Mistral's Generated AI technology allows businesses to launch chatbots, search capabilities, and other AI-driven products. With the ChatGPT chatbot, he surprised many by building a model comparable to the technology developed at OpenAI, a US startup that started the AI boom in 2022. Mistral, named after a powerful wind in France, has quickly become popular by developing more flexible and cost-effective machine learning tools. Some of Europe's largest companies have started using the company's technology, including French car giant Renault and financial services firm BNP Paribas.
The French government is fully supporting Mistral. President Emmanuel Macron called the company an example of “French genius” and invited Mensch to dinner at the Elysée presidential palace. The country's finance minister, Bruno Le Maire, has frequently praised the company, and former French digital minister Cedric Haut is an advisor to Mistral and owns a stake in the startup.
The French government's support signals the growing importance of AI. The United States, France, the United Kingdom, China, Saudi Arabia, and many other countries are seeking to strengthen their domestic capabilities, sparking a technological arms race that will impact trade, foreign policy, and even global supply chains.
Mistral has emerged as Europe's strongest contender in a global battle. But many question whether the company can catch up with its larger competitors in the United States and China and develop a sustainable business model. In addition to the considerable technical challenges of building a successful AI company, the computing power required is staggeringly expensive. (France claims it can meet its energy needs with cheap nuclear power.)
OpenAI raised $13 billion, and another San Francisco company, Anthropic, raised more than $7.3 billion. Mensch said Mistral has raised about 500 million euros ($540 million) and has recurring revenue of “several million”. But in a sign of Mistral's promise, Microsoft took a small stake in February, and Salesforce and chipmaker Nvidia backed the startup.
“This may be one of the best shots we have in Europe,” said Janet Z, managing director of General Catalyst and founding partner of La Familia, the two venture capital firms that invested in Mistral.・Mr. Furstenberg said. “You basically have very powerful technology that unlocks value.”
Mistral agrees that AI software should be open source. This means that anyone should be able to copy, adapt, and reuse programming code. Proponents argue that making the code visible to other researchers would make the system more secure, encourage its use by businesses and governments for uses such as accounting, customer service, and database searches, and improve economic growth. claims to be promoted. This week, Mistral released the latest version of its model online for anyone to download.
In contrast, OpenAI and Anthropic have kept their platforms closed. Open source is dangerous, they argue. That's because open source can be used for bad purposes, such as spreading disinformation or even creating destructive AI-powered weapons.
Mensch dismissed these concerns as the story of “fear-mongering lobbies” including Google, Microsoft and Amazon, which seek to persuade policymakers to enact rules that crush their rivals. He said he is trying to strengthen his dominance.
The biggest risk of AI, Mensch added, is that it will spark a workplace revolution, eliminating some jobs while creating new ones that require retraining. “It will happen faster than previous revolutions, within two years instead of within 10 years,” he said.
Mensch, who grew up in a family of scientists, was fascinated by computers from an early age and learned to program at age 11. He played video games avidly until he was 15, when he decided he “could do something better.” Along with my time. ” After graduating from two prestigious French universities, Ecole Polytechnique and Ecole Normale Superieure, he became an academic researcher at France's prestigious National Center for Scientific Research in 2020. But he soon pivoted to his AI lab, DeepMind, which was acquired by Google, to learn about the industry and become an entrepreneur.
When ChatGPT arrived in 2022, Mensch teamed up with friends from university and decided they could do the same or better in France. In the company's spacious workspace, an army of sneaker-wearing scientists and programmers are currently busy tapping away at keyboards, coding and typing digital texts culled from the internet. So is his bundle of 19th century French literature, which is no longer subject to copyright. Legal — incorporated into the company's larger language model.
Mensch said he is uncomfortable with Silicon Valley's “very religious” enthusiasm for the concept of artificial general intelligence. Technology leaders like Elon Musk and Sam Altman believe that computers will surpass human cognitive abilities, with potentially dire consequences. .
“The whole AGI rhetoric is about God’s creation,” he said. “I don't believe in God. I'm a strong atheist. That's why I don't believe in AGI.”
A more pressing threat, he said, is the threat posed by American AI giants to cultures around the world.
“These models generate content and shape our cultural understanding of the world,” Mensch said. “And as it turns out, French values and American values differ in subtle but important ways.”
As Mr. Mensch grows in influence, he increasingly calls for less regulation, warning that it hurts innovation. Last fall, France successfully lobbied in Brussels to limit the regulation of open source AI systems in the European Union's new artificial intelligence law, a victory that helped maintain Mistral's rapid pace of development.
“If Mistral becomes a big tech powerhouse, it will be beneficial for Europe as a whole,” said O, a former digital minister who led the lobbying effort.
